DBID: Analogy‐Based DSS for Bidding in Construction

Abstract

This paper presents a decision‐support system (DSS) that aids contractors in preparing competitive bids for building projects. The DSS uses neural networks for markup estimation that derive solutions for new bid situations based on analogy with past projects using information elicited from contractors in Canada and the United States. The neural networks were trained to generalize the projects' knowledge and thus become able to predict the outcomes of a project when fed with the contractor's assessment of various project risks. The proposed DSS is coded in a user‐friendly software called DBID. The software enables the contractor to retrain the neural networks on some of his or her past bid encounters and accordingly adapt the model to his or her own environment. In estimating the optimum markup for a new project, the uncertainty in the contractor's assessment of project risks is accounted for by a sensitivity analysis conducted using the Monte Carlo simulation technique. Such analysis produces a measure of the probability of winning at any desired level of markup. The capabilities of the present DSS are demonstrated through an example application.